P Paper8DeterminationofProperties

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/320067938
Ef๏ฌcient determination of mechanical properties of carbon ๏ฌbre-reinforced
laminated composite panels
Article · January 2017
CITATIONS
READS
0
366
2 authors:
Umar Farooq
Peter Myler
University of Bolton
University of Bolton
26 PUBLICATIONS 93 CITATIONS
55 PUBLICATIONS 524 CITATIONS
SEE PROFILE
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Plastic analysis View project
Ultrafast Finite Element Modelling Method of a CFRP Monocoque Electric Car Chassis View project
All content following this page was uploaded by Umar Farooq on 27 September 2017.
The user has requested enhancement of the downloaded file.
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
EFFICIENT DETERMINATION OF MECHANICAL PROPERTIES OF
CARBON FIBRE-REINFORCED LAMINATED COMPOSITE PANELS
Umar Farooq and Peter Myler
Faculty of Engineering and Advanced Sciences, University of Bolton BL3 5AB, United Kingdom
E-Mail: adalzai3@yahoo.co.uk
ABSTRACT
This paper is concerned with integrating experimental and theoretical methods supported by numerical
simulations to efficiently determine mechanical properties of carbon fibre-reinforced laminated composite panels. Ignition
loss experiments were conducted for eight, sixteen, and twenty-four ply laminates to approximate fibre volume fractions by
weights. Rule of mixture was utilised to approximate basic mechanical properties (Young’s and shear moduli and
Poisson’s ratios) for a lamina. The mechanical properties were utilised to develop coefficients of stiffness and compliance
matrices. The coefficient matrices are used in constitutive equations to align off-axis fibres and applied load to mid-plane
direction in two-dimensional formulations. Based on the two-dimensional formulations stacking sequences of threedimensional laminate layupswere developed without resorting to three-dimensional micro-macro mechanics. The
formulations laminates were then coded in commercial software MATLAB TM to predict mechanical properties. Tensile and
flexural physical tests of the laminates were also conducted to validate the simulation obtained mechanical properties.
Comparisons of mechanical properties have shown good agreement (over 90%) between laminates having different types
of stacking sequences. Based on comparison of the results an efficient and systematic two-dimensional methodology is
proposed to predict mechanical properties of three-dimensional laminates.
Keywords: A. composite laminates, B. mechanical properties, C. tensile test, D. bending test.
1. INTRODUCTION
The fibre-reinforced composite laminates are
being increasingly used in structural design as a building
block of modern commercial aircrafts parts due to their
high specific strength, high specific stiffness, light weight,
and superior in-plane properties [1] and [2]. When a
material is characterised experimentally, engineering
constants are measured instead of the stiffness or the
compliance. This is because engineering constants can be
easily defined and interpreted in terms of simple state of
stress and strain used in design development and analysis.
Mechanical properties of the test standard laminates are
obtained from static testing before using them as a
structural component. But laminates are anisotropic in
nature and have to undergo series of experiments to obtain
their properties. Moreover, parts have certain
characteristics of shape, rigidity, and strengthhence their
testing could be complicated [3]. Knowledge of
mechanical properties to evaluate their performance and
identify load bearing parameters that influence strength
and response of laminates (fibre, matrix and interfaces)
require efficient techniques [4]. In the literature common
existing experimental methods consist of: tensile,
compression, flexural, shear modulus, Iosipescu, and vnotch-rail prevail. Most of the methods are detailed in
(ASTM: D7264). A large number of factors affect
property determination of laminates such as: dispersion
and distribution of matrix and filler, compatibility, nature,
and material processing technology [5] and [6].Hence
series of tests are conducted for screening the properties
before their integration in full scale structures or putting
them into work. At the same time, physical testing of
laminates is very inconvenient and resource consuming.
Interfacial structural element and morphology also affect
quality of the evaluated mechanical properties [7] and [8].
Thus larger span-to-depth ratios are used to reduce the
adverse influence of the interfaces [9] and [10]. Many
experimental test methods use different geometries and
holding-fixtures for laminates that produce different data.
Influence of shear effects in the displacements is another
important factor [11] and [12]. The laminates subjected to
off-axis loading system present tensile-shear interactions
in its plies. The tensile-shear interactions lead to
distortions and local micro-structural damage and failure,
so in order to obtain equal stiffness in all off-axis loading
systems, a composite laminate have to be balanced angle
plies [13]. Some of the studies complement physical
testing by using Rule of Mixture (ROM). The method
could be useful approximate laminates with aligned
reinforcement (stiff along the fibres) and very weak
(transverse to the fibres direction). Furthermore, the ROM
method relates material properties of the laminates into
algebraic set of equations that are easier to code and solve
using computer [14]. Still use of the method is limited for
cases of tensile-shear interaction if the off-axis loading
system does not coincide with the main axes of a single
lamina or if the laminate is not balanced [15]. The solution
to obtain equal stiffness of laminates subjected in all
directions within a plane is presented by various authors
by stacking and bonding together plies with different
fibres orientations [16] and [17]. The allocation of the
appropriate input engineering parameters such as the
effective elastic moduli and the associated Poisson’s
Ratios for the materials based on the theory of linear
elasticity and their limitations are described in [18].
Theoretical approach to compute the elastic constants of
laminates based on epoxy resin reinforced alternatively
with glass, HM carbon, HS carbon and Kevlar49 fibres are
reported in [19] and [20]. The laminates taken into account
the plies sequence [0/45/90/-45/0] in order to obtain equal
1375
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
stiffness in all loading systemssubjected to off-axis
loading systems. The elastic constants were also
determined for laminates have to present balanced angle
plies [21].
Review of the literature revealed that most of the
existing studies are experimental where many test methods
have limitations with testing and data logging systems.
Majority of the analytical studies are based on ROM
methods and limited to simplified two-dimensional
formulation.
In the present study, effective mechanical
properties for three-dimensional laminates were
approximated utilising two-dimensional laminated plate
theory. Quantities of volume fractions by weight were
determined with ignition loss method. The volume
fractions were utilised in ROM relations to approximate
basic mechanical properties. The mechanical properties
were utilised in laminated plate formulation to include
load-deformation influence. The formulation was then
coded in MATLABTM to approximate the mechanical
properties. Tensile and flexural tests were also conducted
to compare and validate the results. Comparisons of
experimental and simulation results were found within
acceptable agreement. The study demonstrated that two-
dimensional laminated plate formulation can be
systematically and effectively utilised to evaluate a range
of engineering constants of three-dimensional laminates.
2. MATERIAL AND METHODS
2.1 Geometric properties of the laminates
Composites are heterogeneous materials hence
full characterisation of their properties is difficult as
various processing factors can influence the properties
such as misaligned fibres, fibre damage, non-uniform
curing, cracks, voids and residual stresses. These factors
are assumed to be negligible when care is taken in the
manufacturing process. It is for this reason that the
purpose specific aerospace specialist fabricated laminates
(manufacturers’ supplied) were used here. For a better
understanding of the laminate, a brief illustration of the
coordinate systems used is shown in Figure-1. The X-Y-Z
system is the global coordinates system. The 1-2-3 system
is the local material coordinates system defined foreach
ply with the 1 axis representing the fibre direction, the 2
axis representing the direction perpendicular to the fibre
direction and the 3 axis representing the out-of-plane
direction.
Figure-1. Coordinate systems convention.
A typical 8-Ply laminate is shown in Figure-2.
a) Fibre angle
b) Satin-weave property
Figure-2. Schematic of 8-Ply symmetric laminate.
The laminates are also assumed to be void- free,
linear elastic with plane dimensions 150mm x 120mm
with fibre horns technique of every fourth layer. Average
thicknesses of the laminates consisting of eight-, sixteen-,
and twenty-four plies with layup sequence are given in
Table-1.
1376
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Table-1. Measured thickness of laminates.
Laminates
Code: Fibredux 914C-833-40
No. of ply
Lay-up code
Average thickness mm
8
[0/90/45/-45]S
2.4
16
[0/90/45/-45]2S
4.8
24
[0/90/45/-45]3S
7.2
The material properties are given in Table-2.
Table-2. Material properties of the laminates.
Property parameters
Tensile Modulus Exx,
Eyy in 00& 900 and
in 450& -450
Shear Modulus G12
Units
Fibredux
914C-833-40
Gpa
230
Gpa
23
Gpa
88
Poisson’s Ratio (๏ฎ12)
0.21
2.2 Volume fractions obtained from ignition loss
method
Densities of fibres and resin–matrix contents
were determined experimentally by weighing them in air.
The ignition loss method (ASTM: D2854-68) was used for
polymeric matrix composites containing fibres that do not
lose weight at high temperature. In this method, cured
resin is burnt off from a small test at 5850C in a muffle
furnace. After burning for three hours at 585 0C the density
of fibres comes down to 1.8g/cm3 and that of the matrix as
2.09g/cm3. Three laminates were tested and fibre
percentages (residue mass/sample mass) per unit volume
(cm3) were calculated as shown in Table-3 given below.
The density quantities shown in column 7 were then
utilised in ‘Rule of Mixture’ to calculate volume fraction
of carbon fibresas shown in column 8 of the same Table.
Table-3. Fibre contents of laminates.
Length
Width
Depth
8-Ply
22.40
2.13
5.89
Heated up to 585 0C
Sample
Residue
mass g
mass g
0.4434
0.17
16-Ply
13.4
5.84
11.56
1.4555
0.7396
1.816
50.8
16-Ply
12.61
5.75
11.47
1.3924
0.7027
1.786
50.5
Sample
3.
METHODOLOGY
TO
MECHANICAL PROPERTIES
DETERMINE
Conventions used in this study:
a) Elastic constants are expressed throughout in normal
format rather than italic to keep uniformity with the
other variables in equations.
b) Micromechanics does not refer to mechanical
behavior at the molecular level rather it looks at
components of a composite (matrix and fibre) and
tries to predict the behavior of the assumed
homogeneous composite material.
c)
Density
g/cm3
1.779
Carbon fibre %
50.7
3.1 Micro-mechanics methods
A lamina (heterogeneous at the constituent level)
forms the building block of laminated composites based
structures. The mechanical and physical properties of the
lamiae (reinforcement and matrix) and their interactions
are examined on a microscopic level on various degrees of
simplifications. Fibre and matrix densities are measured
and converted into respective fibre volume fractions that
form basis for approximation of engineering properties
(see Figure-3).
The behavior of the lamina is called “macromechanics”. Most of the structural parts use laminates
that consist of several plies with different orientations
connected together through a bonding interface.
1377
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
ρc Vc = ρf Vf + ρ V (Rule of mixture for density)
(6)
Eq. (1) can be rearranged as
Vc =
−
W
Wc −Wf
( f)+
ρf
Wc
ρc
Where:
The measured densities and fibre volume
fractions of a laminate can be related to its ingredients
using ‘Rule of Mixture’ and utilised to determine volume
fractions by weight. The volume and weight fractions are
given in Eq. (1) and Eq. (2) below:
m
c
Where: Vf =
= composite volume
f
= matrix volume fraction; and V =
fraction
v
c
= void volume
The weight fractions are given with the relation:
Wc = Wf + W = composite weight (void weight is
neglected)
Where: Wf =
m
W =
c
f
c
ρc = Wf
c
=
=
ρf
W
+ m
f
ρf
W
ρm
+
f
= fibre weight fraction and
) is used to approximate
(3)
c−
ρm
Wc −Wf
ρm
Wc
ρc
( ρ f )+
(2)
= matrix weight fraction
The density (ρ =
volume fractions = density
ρc
(1)
=fibre volume fraction; V =
c
f
(7)
The symbols W, V, and ρ represent weight,
volume, and density. The subscripts c, f, m, and v denote
composite, fibre, matrix, and void, respectively.
Fibre volume fraction from densities neglecting
void contents zero is:
Figure-3. Micro and macro-mechanics processfor
property determination.
Vc = Vf + V + V =
ρm
(4)
(5)
FVF =
ρc −ρm
(8)
ρf −ρm
The fibre volume fraction from fibre weight
fraction
FVF = [ +
Where
๐‘Š =
ρF
−
ρm
ρF
[ρm + (๐œŒ๐‘“ −๐œŒ๐‘š ). ๐‘‰
−
]
(9)
(10)
]
The cured ply thickness from ply weight
(www.gurit.com):
CPT
Where:
FVF
FWF
ρc
ρm
=
ρf
WF
CPT (mm)
F
(11)
ρf
= fibre volume fraction
= fibre weight fraction
= density of composite (g/cm3)
= density of cured resin/hardener matrix
(g/cm3)
= density of fibres (g/cm3)
= fibre area weight of each ply (g/m2)
= Cured ply thickness calculated from
volume fraction.
From the Eq. (9) and Eq. (10) fibre volume
fraction by weight was determined as 54% using
MATLABTM code Figure-4.
1378
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
% Computing values from mat lab
% Effective Length = Length 120 mm – Grips 30 mm = 90 mm
l =90;b =23; delta =0.4;
ply=input('enter no of ply');
if(ply==8)% p is applied load
p =25; h =2.4;
elseif (ply==16)
p=190; h=4.8
else (ply==24)
p=600;h = 7.2;
end
mom =p*l/2; i= 23*(h^3)/12; smax=(mom*(h/2))/i; elast=(p*l^3)/(48*delta*i; epsi=smax/elast;
disp(mom);disp(i);disp(smax);disp(elast);disp(epsi)
% Values computed from Matlab software for sample D
% Average length
l1= 12.84;l2=12.38;l=(l1+l2)/2;
%Average width
w1= 11.4;w2=11.54;w= (w1+w2)/2;
%Average thickness
t1= 6.84;t2=4.68;t=t1+t2;
% Volume of sample D
volume = l*w*t;
disp('volume of the sample is'); disp(volume);
% sample mass and density
smass = 0.3672;denst1=smass/volume;
disp('density of the sample is'); disp(denst1);
% Residue mass and density
residu = 0.1666; rmass = 0.1666; denst2 = rmass/volume;
disp('density of the residue is ');disp(denst2)
% Calculations for volume and weight fibre fractions
imass = 1.3924; % initial mass fmass = 0.7027; % final mass
ratio1 =fmass/imass; ratio1= 1/ratio1; ratio1=ratio1-1
% values found from the Internet
rowf = 1.8; % fibre density rowm = 2.09; % matrix density
ratio2=rowf/rowm; denom1 = ratio1*ratio2; denom1 = denom1+1;
fvf = 1/denom1;
disp (fvf)= 0.5419
% fibre weight fraction from fibre volume fraction
rowf= 1.8;rowm=2.09;fvf = 0.5419;
denom1 = rowm+(rowf-rowm)*fvf;fwf=rowf*fvf/denom1;
disp(fwf)
=0.5207
Figure-4. MATLABTM code to compute volume fractions by weight.
3.1.1 Mechanical properties of a lamina
By the rule of mixtures, the modulus of a
composite is defined as the combination of the modulus of
the fibre and the modulus of the matrix that are related to
the volume fractions of the constituent materials [11]:
Ec = Ef Vf + E V
(12)
Where: Ec is the modulus of elasticity of the
composite Assuming perfect bonding very small modulus
of matrix, the equations for calculating ply moduli are
written as:
E = Ef Vf + E V (Longitudinal Young’s modulus)
(13)
E =
G
ν
m f+ f m
m f
f m
=
m f+ f m
(Transverse Young’s modulus)
(14)
(In-plane shear modulus)
(15)
= νf Vf + ν V (Poisson’s ratios)
(16)
3.2.1 Determination of critical volume fractions
Fibre volume fraction using Eq. (9) and Eq. (10)
can be re-written as:
Vf =
f
ρf
[
f
ρf
+
−
− Wf ρ ]
(17)
1379
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
ρc = x [
f
ρf
+
− Wf ρ ]
−
(18)
In terms of fibre volume fraction,๐•๐Ÿ , the
composite density, ๐›’๐œ , can be written as
ρc = ρf Vf + ρ
− Vf
(19)
From the assumption of perfect bonding between
fibres and matrix, we can write:
εf = ε = εc
(20)
σf = Ef εf = Ef εc
(21)
Since both fibres and matrix are elastic, the
representative longitudinal systems can be calculated as
σ = E ε = E εc
(22)
Comparing Eq. (21) and Eq. (22) and knowing
from material properties that Ef โ‰ซ E we conclude that
the fibre stress σf > σ . The total tensile load P applied
on the composite lamina is shared by fibres and matrix so
that
P = Pf + P
(23)
Equation (23) can be written as σc Ac = σf Af +
σ A (24)
or
σc = σf
Af
Ac
Since๐•๐Ÿ =
+σ
๐€f
๐€c
Am
Ac
and ๐•๐ฆ =
σc = σf Vf + σ V
=σf Vf + σ
(25)
๐€m
๐€c
, the Eq. (25) gives
− Vf
(26)
(27)
Dividing both sides of the Eq. (27) by εc and
using Eq. (21) & (22), we can write the longitudinal
modulus for the composites as
E = Ef Vf + E
− Vf
(28)
Equation (28) is called the rule of mixture that shows that
the composite’s longitudinal modulus is intermediate
between fibre and matrix moduli. The fraction of load
carried by fibres in a unidirectional continuous fibre
lamina is
Pf
= σf Vf [σf Vf + σ
− Vf ]−
P
= Ef Vf [Ef Vf + E
− Vf ]−
(29)
In general, fibre failure strain is lower than the
matrix failure strain. Assuming all fibres have the same
strength, the tensile rupture of fibres will determine the
rupture in the composite. In polymeric matrix composite
๐„
is f > . Thus, even for ๐•๐Ÿ = . , fibres carry more than
m
70% of composite load. Thus, using Eq. (28), the
longitudinal tensile strength σL of a unidirectional
continuous fibre can be estimated as
σL
= σf Vf + σ
− Vf
(30)
Where: σf is fibre tensile strength; σ is matrix stress at
fibre failurestrain (ε = εf ).
For effective measurement of the reinforcement
of the matrix (๐›”๐‹
๐›”๐ฆ , the fibre volume fraction in
the composite must be greater than a critical value. This
critical volume fraction is calculated by setting ๐›”๐‹ =
๐›”๐ฆ . Thus, from Eq. (30), the value can be calculated as:
Vf_C
i ica
= [σ
−σ
][σf − σ
]−
(31)
The fibre weight fraction was computed as 50%
using the computer code given in Figure-4 agrees with the
major requirement of the acceptable range of (±5%) fibre
contents to determine elastic constants before using them
in the investigation.
3.1.3 Formulations for coefficients of stiffness and
compliance matrices
The relationships of composite material
properties, relative volume contents, and geometric
arrangement of the constituent materials are useful in
mechanics of materials models. Engineering constants
measured from rule of mixture are used to determine
components of lamina the stiffness matrix [ ] and
compliance matrix [ ]− from Eq. (13)-(16):
=
=
=
=
=
=
−
−
−
−
=
=
−
=
=
=
(32)
−
The
five
engineering
constants
in
stiffness/compliance matrices are very useful in the
analyses of laminates having multiple laminae in nonprincipal
coordinates.
Simple
relationships
of
transformation of stress components between coordinate
local-global axes for the wedged-shape differential
element can be applied to the equations of static
equilibrium of loaded laminates.
1380
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
3.2 Macro-mechanics methods
3.2.1 Formulations for orthotropic lamina
The ‘macro-mechanical’ stress-strain relations of
the lamina can be expressed in terms of an equivalent
homogeneous material. However, the properties of the
composites are usually anisotropic. In the angle-ply
laminates the principal directions of the orthotropy of each
individual ply do not coincide with the generalised
coordinate system. Components of the lamina stiffness
matrix need to be transformed into a global form with
different angles. A unidirectional composite has three
mutually orthogonal planes of material property symmetry
(i.e., the 12, 23, and 13 planes) and is the orthotropic
material. The 123 coordinates are referred as the principal
material coordinates since they are associated with
reinforcement directions. The lamina is in a twodimensional state of stress (plane stress). The stress-strain
relationships can be simplified by letting out-of-plane
shear stresses zero (๐œŽ =
=
=
of a thin elastic
lamina.The elastic constants influence from in-plane
deformations have strain-stress relationships as:
= (๐œŽ
− ๐œŽ )
= (
)
(− ๐œŽ
=
+๐œŽ )
(33)
๐œŽ
๐œŽ
{ ๐œŽ } = [๐‘‡] { ๐œŽ }
(34)
Where:[๐‘‡] = [
−
cos ๐œƒ ๐‘Ž
= sin ๐œƒ
−
−
];
=
The elastic constants (
and
)in
global coordinates may be determined to relate properties
of a lamina in which continuous fibres are aligned at angle
๐œƒ as shown in Figure-5.
=
ν
=
c
θ
i
θ
+
=E [
i
+
+
ν
ν
c
θ
θ
+
−
+
[
−
ν
+
[
−
ν
−
+
=
+
ν
ν
+
+
] sin
(35)
] sin
(36)
−
−
cos
sin
]
ν
(39)
Figure-5. Unidirectional lamina reinforced with
rotated fibres.
There is no coupling between the shear stresses
and normal stress for an orthotropic. The strain-stress
relations for a lamina in a plane stress form become:
=
Stresses in the xy-coordinate system can be
transformed for an orthotropic lamina using the localglobal coordinate transformation matrix:
=
ν
๐œŽ
=−
๐œŽ
−
๐œŽ
=−
๐œŽ
๐œŽ
+
−
−
๐œŽ
−
+
(40)
Coefficients of mutual influence (
and
) for
an angle-lamina in global coordinates can be determined
from the equations:
๐œˆ
= sin ๐œƒ [
−
]
−
]
=
=
=
=
= sin ๐œƒ [
๐œˆ
+
−
−
๐œƒ
+
๐œˆ
+
−
−
๐œƒ
+
๐œˆ
+
(41)
+
(42)
The symmetry presented by the unidirectional
lamina makes it so-called orthotropic material. For an
especially orthotropic lamina (๐œƒ = and
), the stressstrain relations yield:
=
๐œŽ
=−
−
๐œŽ
=
๐œŽ
+
๐œŽ
(43)
(37)
Using Eq. (32), the relations may be written in
local coordinates as:
(38)
๐œŽ
{๐œŽ } = [
] {๐›พ }
(44)
1381
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Similarly, for the general orthotropic lamina
(๐œƒ ≠ and
), the complete set of transformation
equations for the stresses in the xy-coordinate system can
be developed using the local-global coordinate
transformation matrix. The generally orthotropic laminate
creates fully populated, the reduced transformed stiffness
matrix:
ฬ…
๐œŽ
๐œŽ
{ } = [ฬ…
ฬ…
ฬ…
ฬ…
ฬ…
ฬ…
ฬ… ]{
ฬ…
}
(45)
Where matrices: [ ฬ… ] = [๐‘‡]− [ ][ ][๐‘‡][ ]− and Reuter
transforms
[ ]=[
]for strains can be performed in the
same manner using tensor strain as engineering shear
strain. It is not a tensor quantity and is twice the tensor
shear strain (
=
.
The models given in Eq. (45) are complicated
hence semi-empirical models have been developed for the
design purposes. They can be used over a wide range of
elastic properties and fibre volume fractions. The
equations are semi-empirical in nature since involved
parameters in the curve fitting carry physical meaning.
The sharp drop in modulus as the angle changes slightly
from 00is its limitation since over much of the range of
lamina orientation the modulus is very low which requires
transverse reinforcement in most composites. The shearcoupling effects are the generation of shear strains by offaxis normal stresses and the generation of normal strains
by off-axes shear stresses. The degree of shear coupling is
defined by dimensionless shear-coupling ratios or mutual
influence coefficients or shear-coupling coefficients. The
mutual influence coefficients can be found from:
=
,
σ
γ
ε
(46)
Similarly, when the state of stress is defined as
≠ , σ = τ = , the ratio
=
,
γ
ε
(47)
Pure shear stresses τ ≠ , σ = σ =
, the ratio
, characterises the normal strain response
along the y direction due to a shear stress in the x-y plane.
The ratio can be found as:
,
=
τ
ฬ…
(48)
Superposition of loading, stress-strain relations in
terms of elastic constants are:
{
๐‘ 
}= −
−
๐œˆ
๐œˆ
๐œ‚๐‘ 
๐œ‚๐‘ 
๐œ‚๐‘ 
๐œ‚๐‘ 
๐œŽ
๐œŽ
{ }
(49)
๐‘ 
[
]
The modulus of elasticity and tensile strength of a
laminate under a uniaxial load applied in the x-direction at
an angle θ to the fibres 1-direction may be determined
from the transformed reduced stiffness matrix. The
moduli: E , E , G , andν in global coordinates can be
written as:
E
= ⁄
[
E
= ⁄
[
G
= ⁄
[
ν
=
ν
=
⁄E
(
⁄E
(
E
−
+
n −m ν
+
]
n −m ν
+
m −n ν
+
]
+ν
⁄
[
E ν
−
m −n ν
m ν
+
−n
+
−
−m
+
+ν
+
n ν
(50)
(51)
(52)
]
]
⁄E
(53)
(54)
=
⁄G
=[
+ν
−
+ν
(55)
=
⁄G
=[
+ν
−
+ν
+
)
)
]
]
−
(56)
The Eq. (55) and (56) illustrate an angle-ply
laminate in which the various plies are orientated at ± θ to
the plate element axes, in this case the x-axis. The
laminates which have an equal member of + θ and -θ plies
are balanced about their mid-plane are orthotropic in
nature. In the case stresses and strains related by the
transformed reduced stiffness matrix:
σ
σ
σ
=
σ
σ
{σ }
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
[Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ…
Q
ฬ… ]
Q
ε
ε
ε
(57)
γ
γ
{γ }
The thirteen constants ฬ… are related to nine
through the following transformation the components
of the transformed stiffness matrix defined as follows:
1382
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
ฬ…
ฬ…
Q
=
= Q
ฬ…
ฬ…
=
=
ฬ…
ฬ…
Q
ฬ…
Q
ฬ…
ฬ…
ฬ…
ฬ…
ฬ…
ฬ…
=
๐œƒ+
+Q − Q
= Q cos
= Q
=
=
=
=
=
=
−Q
−
−
๐œƒ+
+ Q sin
๐œƒ+
๐œƒ+
+
sin cos + Q
๐œƒ
+
๐œƒ
Q
+ Q
− Q sin cos
− Q −Q − Q
−
cos ๐œƒ
๐œƒ
−
−
−
sin ๐œƒ cos ๐œƒ
๐œƒ
−
sin ๐œƒ cos ๐œƒ
๐œƒ+
๐œƒ
+
−
−
๐œƒ+
๐œƒ
๐œƒ+
๐œƒ
cos
sin
cos
cos sin
sin ๐œƒ
๐œƒ
๐œƒ
+ sin
3.2.2 Formulation for effective constants of laminates
The quantities of the elastic constants of every
plies in material coordinates were utilised to approximate
effective elastic constants for laminated structural element
of thickness H made of N plies in global coordinates:
ฬ… =
๐œƒ
๐œƒ+
when subjected to a shear stress. The state of stress is
defined as whereσ ≠ , σ = τ = .
ฬ… =
(58)
Although the transformed - matrix now has the
form as that of anisotropic material with nine nonzero
coefficients, only four of the coefficients are independent
because they can all be expressed in terms of the four
independent stiffness of the specially orthotropic material.
The lamina engineering constants can also be transformed
from principal material axes to the off-axes coordinates.
The effects of lamina orientation on stiffness are difficult
to assess from inspection stiffness transformation
equations. In addition, the eventual incorporation of
lamina stiffness over the laminate thickness, and
integration of such complicated equations is also difficult.
In view of the difficulties, a more convenient form of
lamina stiffness transformation equations has been
proposed in [12]. By using trigonometric identities to
convert from power functions to multiple angle functions
and then using additional mathematical manipulations.
The invariants are simply linear combinations of the Qij,
are invariant to rotations in the plane of the lamina. Thus,
the effects of lamina orientation on stiffness are easier to
interpret. Invariant formulations of lamina compliance
transformations are also orthogonal. From Eq. (29) it
appears that there are six constants that govern the stressstrain behaviour of a lamina. However, the equations are
linear combinations of the four basic elastic constants, and
therefore are not independent. Elements in stiffness
matrices can be expresses in terms of five invariant
properties of the lamina using trigonometric identities in
[21].
The invariants to rotations are simply linear
combinations in plane of the lamina. There are four
independent invariants, just as there are four independent
elastic constants. In all the stiffness expressions (except
coupling) consist of one constant term which varies with
lamina orientations. Thus, the effects of lamina orientation
on stiffness are easier to interpret very useful in computing
elements of these matrices. The element of fibrereinforced composite material with its fibre oriented at
some arbitrary angle exhibits a shear strain when subjected
to a normal stress, and it also exhibits an extensional strain
๐œˆฬ…
๐œˆฬ…
ฬ…
=
=
=
๐‘
๐‘
∑๐‘=
๐œƒ
∑๐‘=
๐‘
๐‘
๐‘
๐œƒ
∑๐‘=
∑๐‘= ๐œˆ
(59)
(60)
๐œƒ
(61)
๐œƒ
∑๐‘= ๐œˆ
(62)
๐œƒ
(63)
The effective elastic constants in the x-axis
direction Ex, in the y-axis direction Ey, the effective
Poisson’s ratios νxy and νyx, and the effective shear
modulus in the x-y plane Gxy are computed. As symmetric
balanced laminates were considered therefore the
following three average laminate stresses were defined [1]:
σ =
σ =
τ
=
∫−
∫−
⁄
⁄
⁄
∫−
⁄
⁄
⁄
σ
σ
τ
๐‘ง
๐‘ง
(64)
(65)
๐‘ง
(66)
Where: H is the thickness of the laminate.
Integration through-thickness in Eq. (64)-(66) with was
approximated as a summation to obtain the average
stresses and related to the force resultants
(N , N , andN as:
σ =
σ =
τ
=
N
(67)
N
(68)
N
(69)
To relate strains to the stresses:
{
๐‘Ž
} = [๐‘Ž
๐‘Ž
๐‘Ž
๐‘Ž
๐œŽฬ…
๐œŽ
]{ ฬ… }
ฬ…
(70)
The effective elastic constants can be obtained for
the laminate using 3 x 3 balanced matrix at RHS in Eq.
(70):
1383
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
ฬ… =
๐‘Ž
ฬ… =
๐œˆฬ…
ฬ…
๐œˆฬ…
(71)
(72)
๐‘Ž
=
๐‘Ž66
=−
=−
(73)
๐‘Ž
(74)
๐‘Ž
(75)
๐‘Ž
๐‘Ž
=
Where
ฬ…
๐œˆ
ฬ…
๐‘Ž
๐‘Ž
๐‘Ž
=
(76)
(78)
−
=
=
(77)
−
=
Where:
The effective Poisson’s ratios ๐œˆฬ… and ๐œˆฬ… are
inter-dependent and related by the following reciprocal
relations:
ฬ…
๐œˆ
ฬ…
๐‘Ž
(79)
−
(80)
66
=∑
=
[ฬ… ] ๐‘ง −๐‘ง
−
,
= , , 6; = , , 6.
4. NUMERICAL RESULTS AND DISCUSSION
The three-dimensional formulations (Equation
(45)-(51), Equation (56)-(60), Equation (59)-(63), and
Equation (71)-(75) were implemented in MATLABTMV
7.10a code to approximate the elastic constants (see
Figure-6). Input properties shown in Table-2 were
assigned to the respective parameters.
clear
clc
%Enter input data
e11=input(‘Enter Ex’, ‘e11’); e22=input(‘Enter Ex’, ‘e22’);
nu12=input(‘Poisson’s ratio’, ‘nu12’); g12=input(‘Shear-modulus’, ‘g12’);
laminate=input(‘Enter no of plies in laminate’, ‘d’);
diary Elastic_moduli_ply8.out
Q = ReducedStiffness(e11, e22, nu12, g12)
Qbar1=Qbar(Q,0);Qbar2=Qbar(Q,90); Qbar3=Qbar(Q,45); Qbar4=Qbar(Q,-45);
Qbar5=Qbar(Q,-45); Qbar6=Qbar(Q,45);Qbar7=Qbar(Q,90);Qbar8=Qbar(Q,0);
z1=-1.2;z2=-0.9;z3=-0.6;z4=-0.3;z5=0.0;z6=0.3;z7=0.6;z8=0.9;z9=1.2;
A=zeros(3,3);
A=Amatrix(A,Qbar1,z1,z2);A=Amatrix(A,Qbar2,z2,z3);A=Amatrix(A,Qbar3,z3,z4);
A=Amatrix(A,Qbar4,z4,z5);A=Amatrix(A,Qbar5,z5,z6);A=Amatrix(A,Qbar6,z6,z7);
A=Amatrix(A,Qbar7,z7,z8);A=Amatrix(A,Qbar8,z8,z9);
a = inv(A);y = 1/(H*a(2,2));H=2.4;
Exx= Ebarx(A,H)
Eyy=Ebary(A,H)
Nu12=NUbarxy(A,H)
Gxy=Gbarxy(A,H)
diary off
function y = ReducedStiffness(E1,E2,NU12,G12)
%This function returns the reduced stiffness matrix size 3 x 3.
NU21 = NU12*E2/E1;
y = [E1/(1-NU12*NU21) NU12*E2/(1-NU12*NU21) 0 ;
NU12*E2/(1-NU12*NU21) E2/(1-NU12*NU21) 0 ; 0 0 G12];
function y = Qbar(Q,theta)
%This function returns the transformed reducedstiffness matrix size 3 x 3.
m = cos(theta*pi/180);n = sin(theta*pi/180);
T = [m*m n*n 2*m*n ; n*n m*m -2*m*n ; -m*n m*n m*m-n*n];
Tinv = [m*m n*n -2*m*n ; n*n m*m 2*m*n ; m*n -m*n m*m-n*n];
y = Tinv*Q*T;
function y = Amatrix(A,Qbar,z1,z2)
%This function returns the matrix after the layer k with stiffness is %assembled.
for i = 1 : 3
for j = 1 : 3
1384
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
A(i,j) = A(i,j) + Qbar(i,j)*(z2-z1);
end
end
y = A;
function y = Ebarx(A,H)
%This function returns the average laminate modulusin the x-direction.
a = inv(A);y = 1/(H*a(1,1));
function y = NUbarxy(A,H)
%This function returns the average laminate Poisson’s ratio NUxy.
a = inv(A);y = -a(1,2)/a(1,1);
function y = Ebary(A,H)
%This function returns the average laminate modulus
a = inv(A); y = 1/(H*a(2,2));
function y = Gbarxy(A,H)
%This function returns the average laminate shear modulus.
a = inv(A);y = 1/(H*a(3,3));
Figure-6. MATLABTM code to compute engineering constants.
The data obtained from the simulations were
plotted against fibres rotated plies in the laminates. The
plots highlight the relation between elastic constants at
various angles. The code for plots of effective values of
four elastic constants as a function of orientation angle in
the range:
๐œƒ ๐œ‹⁄ at the difference of 10 degrees is
shown Figure-7.
clear
clc
%Enter input data
e11=input(‘Enter Ex’, ‘e11’); e22=input(‘Enter Ex’, ‘e22’);
nu12=input(‘Poisson’s ratio’, ‘nu12’); g12=input(‘Shear-modulus’, ‘g12’);
laminate=input(‘Enter no of plies in laminate’, ‘d’);
diary Elastic_Constants.out
fprintf('====== Angle and Elastic constants ======\n');
fprintf('
-----------------------\n\n')
fprintf('Angle \tExx
v12
Eyy
Gxy \n');
fprintf('===== \t======= \t======= \t=======
======== \n');
i=0;
for ii = 0:10:90
i=i+1;
ex1(i) = Ex(e11,e22, nu12, g12, ii);
nuxy(i) = NUxy(e11,e22, nu12, g12, ii);
ey2(i) = Ey(e11,e22, nu12, g12, ii);
nuyx(i) = NUyx(e11,e22, nu12, g12, ii);
gxy(i) = Gxy(e11,e22, nu12, g12, ii);
fprintf('%2d \t\t%5.2f\t\t%5.2f\t\t%5.2f\t\t%5.2f\n',ii, ex1(i),nuxy(i),ey2(i), gxy(i))
disp('Elastic constant')
disp(ex1(i))
plot(ii,ex1(i));
xlabel('Angle, \theta (degree)');
ylabel('E_{xx}(GPa)');
title('Elasitic modulus E_{xx} v Rotation','FontSize',14)
plot(ii,nuxy(i));
xlabel('Angle, \theta (degree)');
ylabel('\nu_{xy}');
title('Elasitic modulus \nu_{xy} v Rotation','FontSize',14);
plot(ii,ey2(i));
xlabel('Angle, \theta (degree)');
ylabel('E_{yy}(GPa)');
title('Elasitic modulus E_{yy} v Rotation','FontSize',14');
plot(ii,gxy(i));
1385
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
xlabel('Angle, \theta (degree)');
ylabel('G_{xy}(GPa)');
title('Shear modulus G_{xy} v Rotation','FontSize',14);
diary off
Figure-7. MATLABTM code to plot engineering constants v rotations.
Plots of the effective engineering constants of
an8-ply laminate as a function of angle orientations in the
range:
๐œƒ ๐œ‹⁄ at the difference of 10 degrees are
selected for discussion. The plot in Figure-8 illustrates
quantities of Young’s modulus in parallel to fibre (x-axis)
directions. Variation of the curve shows maximum value
around 230GPa as expected when fibres were aligned at
zero-directions. It shows severe drop as angle increases
from 00 and this trend continues until angle reaches 900
where value of the modulus drops to around 23GPa.
Changes in quantities of Young’s moduli with respect to
rotations indicate that simulation obtain values are
realistic.
Figure-9. Poisson’s ratios v ply orientation.
Figure-10 illustrates quantities of Young’s
modulus in perpendicular to fibre (y-axis) directions.
Variation of the curve shows minimum value around
23GPa as expected when fibres were aligned at 90 0. It
shows increasing trend as angle increases from 0 0 and this
trend continues until angle reaches 900 where value of the
modulus reaches to around 230GPa. Changes in quantities
of Poisson’s ratio with respect to rotations confirm that
simulation obtain values are realistic.
Figure-8. Yong’s moduli in fibre direction
v ply orientation.
Figure-9 illustrates quantities of Poisson’s ratios.
Variation of the curve shows gradual drop of the values
from 0.2 maximum when fibres are aligned at 0 0 rotations
and minimum around 0.02 fibres were aligned at 900. The
curve shows decreasing trend as angle increases from 0 0
and this trend continues until angle reaches 90 0. Changes
in quantities of Young’s moduli with respect to rotations
support the arguments that simulations obtain values are
realistic.
Figure-10. Young’s moduli in perpendicular to fibres v
ply orientation.
Figure-11 illustrates quantities of shear moduli.
Variation of the curve follows parabolic path. Maximum
values of 88GPa at fibre directions 00 can be seen which
gradually to minimum around 20GPa at fibre direction 45 0
and then reverse trend begins up to 900 where it again
reaches to 88GPa. Such types of variations were expected
and that confirmed that the predicted quantities are
realistic and genuine. Changes in quantities of shearmoduli with respect to rotations also confirm that
simulation obtain values are realistic.
1386
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
5. VALIDATION OF SIMULATION OBTAINED
ENGINEERING CONSTANTS
Figure-11. Shear moduliv ply orientation.
Since elastic constants are interdependent hence
elastic constants were computed for all the laminates. For
the especially orthotropic and transversally isotropic the
constants are G12, G13; E2 = E3; ๏ฎ21 =๏ฎ31, and๏ฎ23 = ๏ฎ32.In
addition, the relationship among the isotropic engineering
constants Eq. (61) is valid associated with the 23 plane, so
that G =
. Effective values for Young’s moduli
5.1 Rule of mixture
The elastic properties were calculated utilising
the respective volume fractions in Eqns. (25)-(28) for 8-,
16-, and 24-Ply laminates. Identical averaged Young’s
moduli E1 and transverse E2were obtained due to quasiisotropic configuration of the laminates with fibre volume
fraction for all the laminates with three different
thicknesses. Quantities of the Poisson’s ratios and in-plane
shear-moduli also computed and good agreement of the
intra-simulated values for every laminate was found. Since
engineering constants are inter-dependent Young’s moduli
suffice the requirements. Hence Young’s moduli
calculated from the volume fraction equations are given in
Table-5.
Table-5. Young’s moduli computed using
‘Rule of Mixture’.
No of ply
Young’s
modulus
GPa
8-Ply
58.5
[0/90/45/-45]S
16-Ply
58.6
[0/90/45/-45]2S
24-Ply
58.8
[0/90/45/-45]3S
Laminate
+๐œˆ
for all the laminates are shown in Table-4.
Table-4. Simulation obtained Young’s moduli.
Laminate
Lay-up
No of ply
Code
Effective Young’s
moduli
GPa
8-Ply
[0/90/45/-45]S
56
16-Ply
[0/90/45/-45]2S
48.6
24-Ply
[0/90/45/-45]3S
45.8
Lay-up
Code
5.2 Hart smith rule
The Hart Smith Rule used to calculate the elastic
modulus of the laminates. The determined elastic constant
for the quasi-isotropic configuration is calculated and
shown in Table-6.
Table-6. Young’s modulus determined from Hart Smith rule.
Supplied elastic modulus 230 GPa comes down to GPa for 54% contents.
Sequence: [45/0/-45/90]
Hart Smith Rule
Angle (degree)
Multiple
Modulus
45
.1
12.
0
1
124.
-45
.1
12.
90
.1
12.
Modulus of the 4-ply quasi-isotropic laminate
(12.+124+12+12)/4
Equivalent to the unidirectional
GPa
5.3 Tensile test experiment
Three laminates from each of the coupon types
were prepared in I-shapes in line with the testing standard
for tensile tests (ASTM: D3039). Average span and width
of each laminate was 120 mm and 20 mm, respectively.
Thicknesses (t = 2.4, 4.8, and 7.2 mm) from each of the
lay-up of 8-, 16-, and 24-Ply varied due to number of plies
within the stacks. Load transfer tabs were adhesively
bonded to the ends of the laminates in order that the load
may be transferred from the grips of the tensile testing
machine to the laminate without damaging the laminate.
Laminates were gripped at both ends. Approximate
dimensions and relevant spans to depth ratios are shown in
Figure-12.
1387
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Figure-12. Schematic of beam laminates with cross-section.
The effective beam length (L) used for all
calculations was length (120 mm) - both the grips (30 mm)
= 90 mm. Average geometrical dimensions can be seen in
Table-7 below.
Table-7. Beams lay-up configuration and geometrical dimensions.
Configuration
Grip: d
Grip: c
Thickness: t
Effective length: L
Mm
[45/90/0/-45]s
3
15
2.4
90
[45/90/0/-45]2s
3
15
4.8
90
[45/90/0/-45]3s
3
15
7.2
90
Photographs of the laminates selected for the tests
are shown in Figure-13(a). The laminates were inserted
within in the fixture holders of the machine by metal grips
as shown in Figure-13(b) and loaded axially at a rate of
1mm/min. The applied tensile load produced tensile
stresses through the holding grips that results elongation of
the laminate in loading direction.
The applied tensile loading and corresponding
voltage output values were recorded by the software
installed on the data acquisition system at specified load
increments so that the curve is plotted for each and every
test data. Measurements for the strains were correlated
from the recorded voltage ranges and scaled by the gain
โˆ†๐‘’
. Where:โˆ† is out-put
from the formula - given =
๐‘ฃ๐พ
voltage (needs to be divided by gain of 341), ๐‘ฃ is bridge
excitation voltage (its value is: 5v), K is the gain-factor
(equal to: 2.13).
Behaviour of the laminate was observed while
load was being applied through variation in out-put
voltage quantities during the tests of every laminate. Most
of the results have shown consistency and linearity for the
longitudinal strain response through system built-in
display screen. A slight nonlinearity in the strain response
was observed due to the influence of the matrix properties.
Measurements of the strains and loads were used to
determine the elastic modulus. Stresses were calculated
from the applied loads on the respective areas. The
recorded strains and calculated stresses were used to
determine the Young’s modulus as shown in Table-8.
Mean moduli were calculated for the every laminate.
Independent tests for every laminate were carried out for
each of the laminates.
Figure-13. Photographs of: a) laminate and b)
INSTRONTM 5585H machine.
1388
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Table-8. Tensile test results.
Laminate
Thickness
Test
mm
T-1
2.4
T-2
T-3
Voltage
(range)
Microstrain
Cross-sectional
area mm2
Load KN
Stress
GPa
Elastic
modulus
GPa
3.14
3458
48
7.6
158.3
45.8
3.1
3414
48
7.4
154.6
45.2
2.6
2863
96
12.50
130.2
46.5
2.7
2973
96
13.
135.4
46.8
2.5
2753
144
16.
111.1
49.4
1.9
20921
144
15.
90.28
53.
4.8
T-4
T-5
7.2
T-6
test, there is no involvement of end-tabs or changes in the
laminate shape. Tests can be conducted on simply
supported beams of constant cross-sectional area. A
schematic of simply supported beam at close to ends flat
rectangular laminate is shown in Figure-14(a). The
laminate is centrally loaded representing three-point
bending test. Dimensions of the three laminates with
cross-sectional areas having constant width but different
thicknesses are shown in Figure-14(b).
5.4 Flexural test experiment
There is a wide variety of test methods available
for flexure testing described (ASTM: D7264) addressing
the particular needs of heterogeneous non-isotropic
materials. In general, flexure test type tests are applicable
to quality control and material selection where
comparative rather than absolute values are required.
Flexural properties of the laminates were determined using
the standard three point bending test method. For flexure
Figure-14. Schematics: a) laminate bending and b) cross-sectional area.
formulation the calculation of Young’s modulus from the
load-displacement relation given in Table-9.
Approximate dimensions and relevant simple
support of span to depth ratios provides base to
Table-9. Formulation used to calculate young’s modulus.
Bending
moment
=
Moment of
area
=
๐‘ค
Max bending stress
๐œŽ
๐‘Ž
=
The machine shown in Figure-13 is used with
changed testing chamber. Two laminates from the three
types of laminates were selected and atypical one is shown
in Figure-15(a) with test chamber in (b). The laminates
were placed in the fixture and gradually loaded. The
laminates were held in place in the machine in flexural test
chamber and gradually loaded. During flexural tests all of
the laminates experienced deflection under increased
=
๐‘ค
Load
=
๐œŽ๐‘ค
Young’s modulus
=
loading. The deflection typically occurred at the middle of
the laminates, which was in contact with the machine
head, therefore considered to be a loaded edge. The
loading produces the maximum bending moment and
maximum stress under the loading nose. The vertical load
could deflect (bend) the laminate and even fracture the
outer fibres of the laminate under excessive loading.
1389
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
The strain values and the other data were
recorded by software installed on the data acquisition
system at specified load increments so that the curve is
plotted for each laminate. In the beginning of the bending
test the load-deflection curves show slightly erratic
behaviour, but along the course of the curves there is a
definable linear portion. Independent tests for every
laminate were carried out and mean moduli were
calculated for the every laminate. Measurements of the
strain values, deflections and loads were used to determine
the Young’s modulus.
As shown in Table-10 below the recorded voltage
range was scaled by gain factor and strains values were
calculated. Stresses were calculated from the recorded
applied load divided by the corresponding cross-sectional
area. Young’s moduli were then calculated using the strain
and stress values and have been shown in column 8 of
Table-10.
Figure-15. a) Test laminate b) test chamber of
INSTRONTM 5585H.
Table-10. Flexural test results.
Test
B-1
B-2
B-3
B-4
B-5
B-6
Laminate
Bridge factor
Micro-strain
Load N
Stress GPa
0.4
440.
20
26.04
Elastic modulus
GPa
52.1
0.39
429.
20
26.04
50.2
0.52
572.
80
26.04
45.5
0.5
550.
80
26.04
45.28
0.54
594.
170
24.59
41.35
0.53
583.
170
24.6
41.13
8-Ply
16-Ply
24-Ply
5.5 Overall comparison of approximated elastic moduli
Selected young’s moduli for all the 8-, 16, and
24-Ply laminates obtained from simulation, rule of
mixture, Hart Smith, tensile, and flexural testing
methodologies are compared in Table-11. Good agreement
of the predicted values of Young’s moduli for every
laminate was found. The comparisons confirmed that
results of the mechanical properties (physical plus elastic
stiffness) delivered from the MATLABTM programs are
reliable.
Table-11. Comparison of Young’s moduli from different methods.
Methodologies applied
Rule of
Hart
Tensile
mixture
Smith
Young’s modulus GPa
Laminate
Simulation
8-Ply
56
58.5
40.4
45.6
52
16-Ply
48.6
58.6
40.1
46.7
45
24-Ply
45.8
58.8
40.
52.2
41
6. CONCLUSIONS
In this investigation, engineering properties of
quasi-isotropic 8-, 16-, and 24-Ply carbon fibre-reinforced
laminated composite panels were determined applying
simulation methodology based on micro-macro mechanics
of the laminates. Physical properties were determined and
ignition loss tests were conducted to evaluate volume
fractions quantities from the rule of mixture. The rule of
mixture and inverse rule of mixture were utilised to
approximate engineering constants. Then two-dimensional
Flexural
stress-strain relation and micro-macro mechanics methods
were applied to formulate lamina, angle lamina, and threedimensional laminate elements. Computer codes were
developed to determine range of the engineering constants.
Selected data obtained from the simulations were verified
with tensile and flexural physical testing. Based on
comparison of the results the following conclusions can be
extracted:
1390
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
๏‚ง
Micro-macro mechanics of a lamina were utilised to
approximate effective elastic constants for threedimensional laminates.
๏‚ง
Simulation produced results were validated against
the values obtained by volume fractions and Hart
Smith rules, tensile, and flexural testing and were
found to be within acceptable range with (±10%)
deviations.
๏‚ง
Elastic constants determined from MATLABTM
simulations were also compared to the intrasimulation values as a function of fibre rotations and
were found to be in good agreement.
Based on comparisons of the results it is
proposed that three-dimensional (full range) of effective
elastic constants can be efficiently determined from
information of the volume fraction, two-dimensional
micro-mechanics laws, and computer simulations with
reduced physical testing.
REFERENCES
[1] Morgan P. 2005. Carbon Fibres and their composites.
Taylor & Francis, Boca Raton, USA.
[2] Yang B, Kozey V, Adanur S, Kumar S. 2000.
Bending, compression, and shear behaviour of woven
glass fibre- epoxy composites. Composites: Part B.
31: 715-21.
[3] Hussain SA, Reddy SB and V, and Reddy VN. 2008.
Prediction of elastic properties of FRP composite
lamina for longitudinal loading. Asian Research
Publishing Network. 3(6): 70-75.
[4] Gao XL, Li K, and Mall S. 2003. A mechanics-ofmaterials model for predicting Young’s modulus of
damaged woven fabric composites, involving three
damage modes, Int J Solids Struct. 40: 981-999.
[5] John EL, XiaowenY, Mark IJ. 2012. Characterisation
of voids in fibre reinforced composite materials.
NDT&E International. 46: 122-127.
[6] Keshavamurthy YC, Nanjundaradhya NV, Ramesh
SS, Kulkarni RS. 2012. Investigation of tensile
properties of fibre reinforced angle ply laminated
composites. Int. J Emerg Technol Adv Eng. 2(4): 700703.
[7] Emilia S, Nicolai B and Paul B. 2001. Mechanical
characteristics of composite materials obtained by
different technologies. Academic J Manufactur Eng.
9(3): 100-105.
[8] Altaf HS, PandurangaduV, Amba PRG. 2013.
Prediction of elastic constants of carbon T300/ epoxy
composite using soft computing. Int J Inn Research in
Sci Eng Technol. 2(7): 2762-2770.
[9] Bradley LR, Bowen CR, McEnaney B, Johnson LR.
2007. Shear properties of a carbon/carbon composite
with non-woven felt and continuous fibre
reinforcement layers. Carbon. 45: 2178-2187.
[10] Mujika F. 2006. On the difference between flexural
moduli obtained by three-point and four-point
bending tests. Poly Testing. 25: 214-220.
[11] Satish KG, Siddeswarappa B, Mohamed KK. 2010.
Characterization of in-plane mechanical properties of
laminated hybrid composites. J Miner Mater
Character Eng. 9(2): 105-114.
[12] Irina P, Cristina M, Constant I. 2013. The
determination of Young modulus for cfrp using three
point bending tests at different span lengths. Scientific
Bulletin-University Politehnica of Bucharest. 75: 121128.
[13] Konrad Gliesche, Tamara Hubner, Holger Orawetz.
2005. Investigations of in-plane shear properties of
±450 carbon/epoxy composites using tensile testing
and optical deformation analysis. Compos Sci
Technol. 65: 163-171.
[14] Bradley LR. 2003. Mechanical testing and modelling
of carbon-carbon composites for aircraft disc brakes.
PhD thesis, University of Bath, UK.
[15] Barbero EJ. 2010. Introduction to composite materials
design, (Second Edition), CRC Press Taylor &
Francis Group.
[16] D. Bhagwan, Agarwal and Lawrence J. 2012.
Broutman, Analysis and Performance of Fibre
Composites. John Wiley and Sons Inc, Unites States.
[17] Romano F., Fiori J., Mercurio U. May 2009.
Structural Design and Test Capability of a CFRP
Aileron. Composite Structures. 88(3): 333-341.
[18] Liyong Tong, Adrian P. Mouritz and Michael K.
2002. Bannister '3D Fibre Reinforced Polymer
Composites' Book, Elsevier Science Ltd. All Rights
Reserved.
1391
VOL. 12, NO. 5, MARCH 2017
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
[19] Prashanth Banakar1, H.K. Shivananda2 and H.B.
Niranjan3. 2012. Influence of Fibre Orientation and
Thickness on Tensile Properties of Laminated
Polymer Composites. Int. J. Pure Appl. Sci. Technol.
9(1): 61-68.
[20] F. Mujika. 2006. On the difference between flexural
moduli obtained by three-point and four-pointbending
tests, Polymer Testing. 25: 214-220.
[21] F. Gibson. 2007. Principles of composite materials
mechanics. Taylor & Francis Group, CRC Press Boca
Raton. pp. 83-126.
1392
View publication stats